Entropy based C4.5-SHO algorithm with information gain optimization in data mining.

Area of Science:

• Computer Science
• Data Mining
• Machine Learning

Background:

• Information efficiency is crucial in IT development and applications.
• Data mining extracts meaningful insights from large datasets for decision-making.
• Existing classification algorithms have limitations in optimizing information gain.

Purpose of the Study:

• To introduce a new classification algorithm for enhanced information management.
• To improve the information gain tuning process in decision tree algorithms.
• To optimize the C4.5 decision tree algorithm using the Selfish Herd Optimization (SHO) algorithm.

Main Methods:

• Combined the classical C4.5 decision tree approach with the Selfish Herd Optimization (SHO) algorithm.
• Tuned the information gain of datasets by updating optimal weights based on SHO.
• Partitioned datasets into two classes using quadratic entropy calculation and information gain.

Main Results:

• The proposed C4.5-SHO method demonstrated optimized decision tree gain.
• Evaluated the robustness of the C4.5-SHO method on various datasets.
• Compared performance against ID3, CART, ant colony optimization, particle swarm optimization, and cuckoo search algorithms.

Conclusions:

• The C4.5-SHO method offers improved information management through optimized decision tree gain.
• The algorithm shows competitive accuracy and area under the ROC curve compared to existing methods.
• This approach contributes to more efficient data mining and decision-making processes.